Adaptively entropy-based weighting classifiers in combination using Dempster-Shafer theory for word sense disambiguation

Abstract

In this paper we introduce an evidential reasoning based framework for weighted combination of classi¯ers for word sense disambiguation (WSD). Within this frame- work, we propose a new way of de¯ning adaptively weights of individual classi- ¯ers based on ambiguity measures associated with their decisions with respect to each particular pattern under classi¯cation, where the ambiguity measure is de¯ned by Shannon's entropy. We then apply the discounting-and-combination scheme in Dempster-Shafer theory of evidence to derive a consensus decision for the classi¯ca- tion task at hand. Experimentally, we conduct two scenarios of combining classi¯ers with the discussed method of weighting. In the ¯rst scenario, each individual clas- si¯er corresponds to a well-known learning algorithm and all of them use the same representation of context regarding the target word to be disambiguated, while in the second scenario the same learning algorithm applied to individual classi¯ers but each of them uses a distinct representation of the target word. These experimental scenarios are tested on English lexical samples of Senseval-2 and Senseval-3 resulting in an improvement in overall accuracy.